Session 2 - EPRImydocs.epri.com/docs/PublicMeetingMaterials/1028/Session_2...Conservation and Optimization Via Volt/var Control Tuesday, October 28, 2014 8:00 a.m. Session Moderators

1 © 2014 Electric Power Research Institute, Inc. All rights reserved. EPRI / DOE The Smart Grid Experience: Applying Results, Reaching Beyond

Session 2: Conservation and Optimization Via Volt/Var Control

The Smart Grid Experience:

Applying Results,

Reaching Beyond

Conservation and Optimization Via Volt/var Control

Tuesday, October 28, 2014

8:00 a.m. Session

Moderators

Jared Green, EPRI

Joe Paladino, DOE

The Smart Grid Experience: Applying Results, Reaching Beyond

October 27-29, 2014


Session Agenda

• (8:05-8:20 AM) Brian Schell – American Electric Power

• (8:20-8:35 AM) Bruce Lovelin – Central Lincoln

• (8:35-8:50 AM) Daniel Fournier ‐ Hydro‐Québec

• (8:50-9:00 AM) 5-10 minute break for Q&A

• (9:00-9:15 AM) Jim Parks – SMUD

• (9:15-9:30 AM) Jay Oliver – Duke Energy

• (9:30-9:40 AM) Jeff Roark – EPRI

• (9:40-9:50 AM) Joe Paladino – DOE

• (9:50-10:00 AM) 5-10 minute break for Q&A


CVR and VVO Basics

Distance from Substation

Vo

lta

ge

126

120

114

Capacitors Capacitors Regulators

Tap-changing

Transformer

Benefits of CVR and VVO:

- Lower losses in lines, transformers.

- Lower losses in end-use devices.

- Allows more real-power throughput.

CVR factor (CVRf) = % change in load ÷ % change in voltage


Brian Schell, American Electric Power Email: [email protected]

• Brian received his B.S.E.E. degree from West Virginia

University in 1988 and has been a licensed

Professional Engineer in the State of Ohio since 1995.

• Principal Engineer employed with American Electric

Power for 25 years.

• Worked first 5 years in the Region Engineering

Organization in Columbus, Ohio.

• Worked for 15 years in the Distribution Planning

organization and was supervisor for Ohio the last few

years.

• Since 2010 worked in the Grid Management

Deployment organization for American Electric Power.


Experience with three vendors: (alphabetical order)

• Cooper Yukon VVO

• GE VVO

• Utilidata AdaptiVolt™ VVO

Volt VAR Optimization at AEP


• In Service

– 17 circuits in Ohio

– 13 circuits in Oklahoma

– 9 circuits in Indiana

• In Progress

– 25 circuits in Kentucky

– 25 circuits in Indiana

– 3 circuits in Michigan

• In Planning

– 80 Circuits in Ohio

– 25 Circuits in Indiana

• Approximately 200 circuits in

the growing VVO Plan



8

Volt VAR Controllers

Mesh Network

Mesh Master

EOL Monitors

Line

Regulators

Line

Capacitors

DMS - GENe

Fiber

or Mesh

Volt VAR Optimization Architecture


Volt VAR Optimization Architecture

DRAFT


116.0

118.0

120.0

122.0

124.0

126.0

0 20 40 60 80 100

Normal Operation

Without VVO = 6-7-11 @ 4:30pm

Substation

EOL CAP 1 CAP 3 REG

Without VVO = 6-7-11 @ 4:30pm

With VVO = 6-6-11 @ 4:30pm

116.0

118.0

120.0

122.0

124.0

126.0

0 20 40 60 80 100

Normal Operation With VVO

CAP 2

AEP Ohio: Gahanna – 4504 (13 KV) Voltage Profile


Circuit Performance In Multiple States


SCADA PI Process Book View


• Volt Var Optimization technology works as expected Testing validates that ~2-4% energy and demand reduction is

achievable.



Demand and Energy Reduction Results

• Measurement &Verification (M&V) Analysis Methods

Requires Day On Day Off Testing

Battelle/PNNL Third Party Analysis

Protocol #1

How do we improve M&V methodology?


Reduced Tap Operations using Utilidata AdaptiVolt

System

Tap Operation Reduction Summary

Average / Day over a year

Phase A – 33.6 reduced to 24.0 (28.6%)

Phase B – 32.6 reduced to 24.8 (23.9%)

Phase B – 32.0 reduced to 22.7 (29.1%)



• Standards (Architecture, Addressing, Regs, CAPs,

LTCs, Sensors, RTUs, etc.)

• Day 2 Support (Resources, Monitoring, Commitment,

Costs, Service Level Agreements, etc.)

• Processes (Alarms, Vendor Updates, Equipment

Failures, Circuit Modeling, etc.)

• Documentation (Version Control, Updates, Central

Repository, etc.)

• Resource Plan (GMD, Telecommunications, Station

Engineering, Distribution Engineering, etc.)

• Testing


Bruce J. Lovelin, Central Lincoln PUD [email protected]

• Chief Engineer and Systems Engineering Manager

• 30 years experience in electric utility industry

• Electrical Engineering Degree from Oregon State

University

• Smart Grid Project Program Manager

Central Lincoln PUD

• 270 MW Peak load, 39,000 Customers

• 130 employees

• 120 miles of Oregon Coastline


Could a AMI System be used to Manage

Customers’ Voltage real time?


Pilot Project

Lincoln Beach Substation • One LTC transformer

• Two feeders

• 1400 meters

• Coastal community

• High percentage of vacation homes

• 8 months heating annually

• No A/C


System Operation


System Operation


System Operation


System Operation


System Operation


System Operation


Project Results

• EDGE Validator uses hourly data • Bus voltage, MW

• Temperature, relative humidity

• CVR on/off

• ON hours paired with similar OFF hours


Lessons Learned

• Great Tool for Finding Service Issues

• Operational Changes

• LTC Operation Unchanged

• SCADA and AMI Integration Went Well

• Did not impact AMI Meter Reading Performance

• No Customer Issues


Project Assessment

• Saved Customers about 2% Energy Consumption

• Operational Positives

• Easy to Implement

• SCADA and AMI Integration Went Well

• Robust Measurement and Verification

• Cost-Effective Resource – 1 cent/kwh


Next Steps

• Full System Implementation

• Up to 25 Substations

• 3 Year Project

• Begins Winter 2015


Thank You!

For a copy of our case study

please contact:

Bruce J Lovelin

[email protected]


Daniel Fournier, Hydro Quebec Email: [email protected]

• Daniel Fournier received a Bachelor in Engineering Physics from the Ecole Polytechnique de Montreal in 1980. He later earned a M.Sc. and Ph.D. in Energy INRS Energy in 1985 and 1988 respectively

• He worked at the Institut de recherche d'Hydro-Québec (IREQ) from 1990 to 2004 and was involved in various research and development related to expert diagnostic systems (infrared thermography and partial discharges) for Distribution equipments

• Since 2004, he worked for Hydro-Québec Distribution as a senior engineer in the field of smart grid applications.

• Since 2010, he has been representing Hydro-Québec Distribution as advisor for the EPRI Smart Grid Demo Host Site Project.

• He is also in charge of the Distribution Automation Telecom modernization project.

• He is a member of the Order of Engineers of Quebec.

October 27-29, 2014

Charlotte, NC

VVC at Hydro-Québec Distribution

by: Daniel Fournier, Eng.

Jordi Drouin, Eng

Laurier Demers, Eng.

The Smart Grid Experience:

Applying Results, Reaching Beyond


Project Description

• Goal: by 2023, 2 TWh savings (annual energy consumption)

• 2008-2010 (7.3M$ invest.) Pilot phase completed

• 2011-2023 (152.4M$ invest.)

150 substations and 2,000 distribution lines

1,000 remotely monitored voltage transformers

865 remote-controlled capacitor banks

Infrastructure upgrades


VVC Project Status

• Voltage control on 17 substations (10 more to come in 2014)

• 195 TTT (medium voltage sensors) installed

• 865 remote controlled capacitors banks installed


Reactive Power Control System status

Overview of system components 865 capacitors at 1.2 Mvar each – total capacity 1038 Mvar

Remote control capacitors are controlled by the distribution control system

576 capacitors currently remote controlled

Objectives Switching of the shunt distribution capacitors by remote control in order to

• improve voltage profile (distribution)

• reduce losses (transmission and distribution)

• increase stability limits (transmission)

• increase the transit of power (transmission)

Consider the real time needs and constraints of the transmission provider

Increase the energy gains if the substation has both volt & var control


Reactive Power Control System

Observations: Pilot project: ~1800 capacitor switching operations

Some issues with capacitor bank operations: mechanical or remote control related deficiencies (approx. 5% total)

Important to use integrated measurements in the algorithm as opposed to instantaneous measurements

No customer complaints related to the reactive power control system observed during the pilot project

Next steps: Improve the algorithm to optimise the number of capacitor operations

Reduce maintenance cycles on the equipment

Increase capacitors being managed by the VAR control system from 576 capacitors to 865 capacitors

Develop indicators to determine the frequency in which each capacitor is used. (less frequently operated units could be moved to another location)

Integrate a VVO system in order to maximize the energy gains


Voltage reduction techniques

1. Permanent setpoint modification

2. Seasonal setpoint modification

3. Dynamic voltage control

4. Dynamic Volt-Var control (VVC)

5. Dynamic voltage control – No online simulator

static voltage targets only

6. Dynamic voltage control – No end-of-line

voltage measurements (simulator only)

Static setpoint

Dynamic setpoint Normal modes

Dynamic setpoint Alternative modes



Summary of requirements

Requirements


Perman. setpoint modif

Seasonal setpoint modif

Dynamic voltage control - Offline

simulator

Dynamic voltage control - No meas.

Dynamic voltage control

Dynamic volt-var control

Offline Simulator (planning tool) x x x

Online Simulator (real time load flow) x x x

Substation data acquisition x x x x

Remotely controlled voltage regulator at

substation x x x x Real time voltage measurement at critical locations x x x

Remotely controlled capacitor banks x



Summary of cost-effectiveness analysis


CVR study on 4 substations (37 lines)


Energy gains (GWh)

1

80

.7

20

1

21

0.5

21

42

.6

8.8

95

16

.5 34

.5

8.1

8.9

0

50

100

150

200

250

2011 2012 2013

Permanent setpoint modifications

Seasonal setpoint modifications

VVC control

Other (day on/day off exp.)


Lessons learned

– Low voltage on substation’s distribution panel:

in certain substations, because of a design flaw, lowering the voltage creates a risk of affecting the operation of the substation itself. Take that into account.

– During planning, two limitations are often overlooked :

• Regulation dead band (tap changer)

• Voltage unbalance (greater than expected). Must use the lowest phase voltage

– Some customers need an adaptation period even though the reduced voltage stays within the standard limits. (ex. hospitals with sensitive equipment).


Conclusion

• It is not realistic to implement VVC in all substations. Still,

there are several other ways of minimizing energy

consumption through voltage reduction.

• Benefits are :

– Allows to minimize cost and/or maximize energy saved

– Could reveal issues before wasting investments.


Next steps:

• Adjust our communication process to keep sensitive customers

informed of every voltage reduction as soon as planned (often

one year ahead)

• Evaluate the cost of fixing the low voltage issue in each

problematic substations

• Evaluate the use of the AMI for voltage measurements

• With the gained knowledge, refine our deployment strategy and

our business case


Reference

• Hydro-Québec Smart Grid Host Site Progress Report

For the Period Ending October 2013 (Technical

Update, February 2014)

EPRI Report: 300200144 (members only report)


Jim Parks,

Sacramento Municipal Utility District Email: [email protected]

• Program manager in the Energy Research and

Development department at SMUD.

• Completing a $308 million smart grid initiative,

SmartSacramento®.

• SmartSacramento has over 40 individual projects.

• Projects include distribution automation, smart

meters, demand response, dynamic pricing and more.

• Past work includes transmission planning, energy

efficiency, electric transportation and emerging

technologies.

• Over 25 years’ experience in the energy industry.


CVR Pilot Project

• Automated 118, 12 kV feeders

• Began summer 2011 on two substations.

– Retrofit of sub controls

– Addition of switched capacitor banks

– Utilization of existing Capcon control system

• Goal of initial phase:

– Test both CVR and VVO.

• Hypothesized that an industry average CVRf (0.5 – 0.7) could be achieved.

• Expanded project in 2014 to 14 substations.

– Wanted to determine operational strategy--peak-period/emergency or 24/7 operation.

Substation A 2% CVR Analysis

Substation

Approximate

Avg. Percentage

Demand

Reduction

(2% V reduction)

Substation A 2.5%

Substation B 1.0%

Early Results


Project Successes - Overview

• Modified control system worked as designed.

• Conducted three years of CVR testing.

• Developed a statistical model to predict CVR

impacts.

– Used a variety of variables, including PV.

– Two separate regression methods produced similar

results.

• 1.8% average voltage reduction ≈

– 2% average daily energy (MWh) reduction

– 1.1 % average load (MW) reduction


Project Successes – 2013 Analysis A

vera

ge C

VR

f

Average Likely Range


Surprises Related to the Project

• Goal was to test CVR at 3% voltage reduction.

– Limited target to 2% (actual average reduction was 1.7%).

119

120

121

122

123

124

125

126

Vo

lts

SUNT1 Voltage Profiles

Day ON Voltage Day OFF Voltage

Day ON Average Day OFF Average




– Limited target to 2% (actual average

reduction was 1.7%).

• “Distribution data is MESSY!”


MESSY Data - Episodic Loads

6

6.5

7

7.5

8

8.5

9

9.5

6:00 AM 6:10 AM 6:20 AM 6:30 AM 6:40 AM 6:50 AM 7:00 AM

MW

SUNT1 Load


MESSY Data – Clearances and Outages

11

11.2

11.4

11.6

11.8

12

12.2

12.4

12.6

1

2

3

4

5

6

7

8

9

1 4 7 10 13 16 19 22 25 28 31 3 6 9 12 15 18 21 24 27 30

KV

MW

Day

CAMC1 Summer 2013 (Aug - Sept)

Substation Load Bus Voltage

RTU

ISSUES NO CHANGE IN VOLTAGE

CLEARANCE

or OUTAGE




– Limited target to 2% (actual average

reduction was 1.7%).

• “Distribution data is MESSY!”

– Episodic loads

– Power transformer clearance for

maintenance.

– Communication failures

• Difficulty explaining the CVRf outliers.

• No known customer complaints.


Reaching Beyond

• Challenge is accurately measuring CVR impacts

– Impact of voltage reduction is small and variable.

– Normal variation in load is comparatively large.

– Small moment-to-moment variations may be larger than CVR impact.


Reaching Beyond


– Impact of voltage reduction is small and variable.

– Normal variation in load is comparatively large.

– Small moment-to-moment variations may be larger than CVR impact.

• Model provides good approximation.


Reaching Beyond


• Model provides good approximation.

• Performed a CBA on the 14 substations in the project.

$ -

$ 200

$ 400

$ 600

$ 800

$ 1,000

$ 1,200

$ 1,400

$ 1,600

$ 1,800

1 5 10 15 20 25 30

Year

thousand

Annual Savings

Accumulated

PV Savings


Reaching Beyond – Next Steps

• Challenge is accurately

measuring CVR impacts

• Developed model provides

good approximation.

• Performed a CBA on the 14

substations in the project.

• Developing a business case

for peak-time and full-time

CVR.

• Present findings to

management.


Jay Oliver, Duke Energy Email: [email protected]

• Held Director of Grid Automation role since Fall 2012

• Prior to this, served as

– Major Projects Manager, Distribution System Demand

Response, Progress Energy Florida

– Director, Distribution Services, Progress Energy Florida

– General Manager, South Coastal Region, Progress

Energy Florida

• Received bachelor’s degree in electrical engineering from

the Georgia Institute of Technology

• Earned his master’s degree in business administration from

the University of South Florida.

• Licensed Professional Engineer in Florida.


IVVC Projects at Duke Energy

Duke Energy Ohio

• IVVC designed to operate 24/7

• Targets an average 2% voltage reduction on approx. 500 circuits

• Included gaining control of voltage regulating devices (in substations

and on distribution circuits) through DMS

Duke Energy Progress

• Distribution System Demand Response (DSDR) targeted 310MW of

peak demand reduction capability to avoid new CT plant construction

• Peak shaving voltage reduction currently approaches 4%

• North Carolina Utility Commission classified DSDR as an Energy

Efficiency program with rider recovery

• Included significant circuit conditioning and automation of voltage

regulating devices through DMS


IVVC Projects at Duke Energy

Other Jurisdictions

(Carolinas, Florida, Indiana and Kentucky)

• Evaluating Volt/VAR control opportunities to

determine appropriate methodologies and

optimal design for each area

• Applying lessons learned and best practices

from IVVC and DSDR projects


IVVC Value Proposition

• Capital investments in IVVC leads to net reduction in

customer bills through fuel savings

.

IVVC Business Rationale for Duke Energy

• Customers win through fuel savings

and net lower bill

• Shareholders win through

increased earnings on investment

capital

http://www.getmoreclientsacademy.com/wp-content/uploads/2012/01/stick_figure_money_puzzle_key_lock_pc_400_clr.png


IVVC Business Rationale for Duke Energy

Key Observation

• IVVC pays for itself today

• IVVC also has multi-purpose capability as a Grid Management tool

that positions the utility for the complexities of the future

Source: EPRI


Integrated Volt/VAR Control – Duke Energy Progress DSDR

* DSDR load reduction capabilities

nearly match the generation

capabilities of Units 3 and 4 of

DE’s Asheville Plant which are

capable of producing a total of

324 megawatts of peaking

power

IVVC Project Results and Benefits - DEP

Completed in July 2014

Load Reduction Capabilities - 310 MW

Peak shaving benefit confirmed

Realizing back stand and spinning reserve benefit

Comparable to dispatching a CT plant*


Growing Customer Load

Trajectory of Load

(dotted)

Sys

tem

Lo

ad

(D

EP

)

System Load for 4 Hours During 2014 Polar Vortex

Generating Limit

IVVC

Load

Reduction

5 A

M

6 A

M

7 A

M

8 A

M

9 A

M

10 A

M

88 Thousand Customer Outages Avoided during extreme cold weather event

IVVC Project Results and Benefits - DEP


IVVC Circuit Commissioning Plan/Actuals

Distribution System Efficiency Metrics - IVVC

Number of Circuits Complete with IVVC by Year

2013 2014 2015 Planned Actual Planned Actual Planned Actual August 50 50 125 162 122 - EOY 100 87 275 20 - - Total 150 137 412 319 534 - Total % 28% 26% 77% 60% 100%

Average System Voltage Baseline (2012) 123.2 V

IVVC Operation (as of 8/31/14)

Avg Circ Volt

w/ IVVC Avg Circ Volt

Red % w/ IVVC MWh Red

w/ IVVC Assumed

CVR Factor # of Cir w/

IVVC

121.1 V 1.75% 33,241 0.5 319

IVVC Project Results and Benefits – DEO

As of 8/31/14


Successes, Surprises, and Next Steps

Lessons Learned

• IVVC projects require dedicated and cross functional project team

• Somebody needs to be “in charge” of delivering the benefit

• Do not underestimate efforts for DMS configuration and testing

• Legacy data accuracy is critical to DMS system model

• Utility and vendor senior management sponsorship is critical

• Received higher level of regulatory oversight than we expected

Next Steps

• Will consider a combination of 24/7 and peak shaving modes

• Developing Duke Energy IVVC road map


Jurisdiction 2015 (year 1)

2016 (year 2)

2017 (year 3)

2018 (year 4)

2019 (year 5)

2020 (year 6)

DEO

DEI

DEK

DEC

DEP

DEF

Foundational

Install

Complete

Foundational

IVVO Build

Initial start

to partial

Final

Benefits

Delivery

Foundational

IVVO Build

Foundational

IVVO Build

2021 (year 7)

Foundational

IVVO Build

Test

Year

Test

Year

Lessons learned; continued

evolution of capability

Lessons learned; continued

evolution of capability

IVVC Conceptual Deployment Timeline

Benefits

Delivery

Foundational

Install

Complete


Jeffrey D. Roark, EPRI Email: [email protected]

• 34 years experience in regulated, unregulated, and

government utilities

– transmission and generation system planning,

– strategic planning

– bulk power contracts

– power market analysis

– wholesale deal structuring

– trading and marketing research

– regulatory analysis (as both regulator and regulated)

• With EPRI since 2011, responsible for cost/benefit

analysis in the Smart Grid Demonstrations program

• BEE, MSEE Auburn University

MBA, University of Alabama at Birmingham


Project: Cost/Benefit Analysis tool

• Developed an Excel-based cost/benefit analysis screening tool for CVR and DA/FLISR.

• Helps analysis of prospective designs; but not a design tool.

• Provides platform for calculating sensitivities to major project variables.


Project Successes

• Applied prototype in distribution automation system build-out study

• Used tool to analyze a number of CVR projects

• Version 1.1 published and available on Epri.com Product ID 3002001294


Cost/Benefit Analysis for VVC/CVR

• CBA for CVR is mostly straightforward.

– Utility Costs: Capital for VVC equipment & systems

– Utility Savings: energy, GT&D capital deferrals

• Capacity and energy impacts are easily monetized.

• The tool handles the largest cost and impact components.


Inputs: Equipment Capital & O&M (Example Numbers)


$ -

$ 50

$ 100

$ 150

$ 200

$ 250

$ 300

1 5 10 15 20 25 30

Year

Thousand

Calculates Revenue Requirements

For CVR Equipment

Replacement

of 10-year

Assets

Replacement

of 15-year

Assets

Replacement of

10 yr and 20 yr

Assets

Replacement of

10 yr and 15 yr

Assets


Inputs: System and CVR-Impact Data (Example Numbers)


Example Case:

Impacts and Avoided Costs

$ -

$ 100

$ 200

$ 300

$ 400

$ 500

$ 600

$ 700

1 5 10 15 20 25 30

Year

thousand

First year of

avoided capacity


$ -

$ 200

$ 400

$ 600

$ 800

$ 1,000

$ 1,200

$ 1,400

$ 1,600

$ 1,800

1 5 10 15 20 25 30

Year

thousand

Example Case: Net Benefits

Annual Savings

Accumulated

PV Savings


Example Case: CVRf Sensitivity

0.3

0.4

0.5

0.6

0.7

0.8

0.9

CVRf = 1

$ (500)

$ -

$ 500

$ 1,000

$ 1,500

$ 2,000

$ 2,500

$ 3,000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

thousand

Accumulated PV of Net Benefits for Various CVRf Values



Not so surprising:

• Utilities’ monetization assumptions (marginal energy and

capacity) may not fit simple inflation scheme.

– Used modified version to incorporate custom marginal

energy and marginal capacity values

• Sensitivities provide sense of assumption importance.

– CVRf variation low to high

– CVRf decay to a lower value


A Few Comments on M&V for CVR

• Low or nonsensical results from day-on/off regression

analysis does not necessarily mean that CVR is not

effective on a particular circuit.

• Large, non-conforming loads or singular events can mask

the CVR effect.

• Evaluate statistical significance of all results, including

high and “expected” results.

• Given M&V issues, nothing is precise.

Look at a range of possibilities.


Questions / Discussion

Documents

Session 2 - EPRImydocs.epri.com/docs/PublicMeetingMaterials/1028/Session_2...Conservation and Optimization Via Volt/var Control Tuesday, October 28, 2014 8:00 a.m. Session Moderators